eScholarship provides open access, scholarly publishingservices to the University of California and delivers a dynamicresearch platform to scholars worldwide.

Department of Economics, UCSBUC Santa Barbara

Title:What do OECD countries cut first at a time of fiscal adjustments? A dynamic panel data approach

Author:Sanz, Ismael, Universidad Complutense de MadridVelazquez, Francisco J, Universidad Complutense de Madrid

Publication Date:12-01-2003

Series:Departmental Working Papers

Publication Info:Departmental Working Papers, Department of Economics, UCSB, UC Santa Barbara

Permalink:http://www.escholarship.org/uc/item/4j744960

Keywords:Fiscal consolidation, Composition of government expenditure, Dynamic panel data, voter groupdecision model

Abstract:Following the present atmosphere of budgetary cuts we analyze the effects of fiscal consolidationon the composition of government expenditures by functions. Using a dynamic voter groupdecision model and exploiting the panel structure of the dataset – 26 OECD countries over theperiod 1970-1997- by GMM estimation we find that fiscal adjustments protect social expenditure.Nevertheless, we find that among social expenditure, those being also productive such aseducation and health are even more protected than those not being productive as socialwelfare expenditure. Moreover, fiscal adjustments do not fall primarily over the most productiveexpenditure, transport and communications, but over defense and economic services.

1

What do OECD countries cut first at a time of fiscal adjustments?

A dynamic panel data approach.

Ismael Sanz

European Economy Group –Universidad Complutense de Madrid Pozuelo de Alarcon, Madrid 28223.

Francisco J. Velázquez Universidad Complutense de Madrid Pozuelo de Alarcon, Madrid 28223.

December, 2003

(*) Francisco Javier Velazquez (javel@ccee.ucm.es) Despacho 106. Pabellon de Segundo Facultad de Ciencias Economicas y Empresariales Campus de Somosaguas. Universidad Complutense de Madrid Pozuelo de Alarcon. 28223 Madrid. Spain Acknowledgements We are grateful to Theodore Bergstrom, Norman Gemmell, Richard Kneller, John

Ashworth, Carmela Martin and Andreas Stephan for their helpful and interesting

comments and suggestions on this paper. Part of this research was undertaken while the

first author was a Visiting Research Fellow of the School of Economics of the University

of Nottingham and the Economics Department of the University of California, Santa

Barbara. We are grateful for the hospitality received and all the research facilities

provided. Financial support from Bank of Spain (Scholarship for Visiting Abroad

Universities) and Science and Technology Minister of Spain (CICYT project, SEC2000-

0751-C03-01) is gratefully acknowledged.

2

What do OECD countries cut first at a time of fiscal adjustments?

A dynamic panel data approach.

Abstract

Following the present atmosphere of budgetary cuts we analyze the effects of fiscal

consolidation on the composition of government expenditures by functions. Using a

dynamic voter group decision model and exploiting the panel structure of the dataset – 26

OECD countries over the period 1970-1997- by GMM estimation we find that fiscal

adjustments protect social expenditure. Nevertheless, we find that among social expenditure,

those being also productive such as education and health are even more protected than those

not being productive as social welfare expenditure. Moreover, fiscal adjustments do not fall

primarily over the most productive expenditure, transport and communications, but over

defense and economic services.

JEL Classification: H10, H50, C23, D72. Keywords: Fiscal consolidation, Composition of government expenditure, Dynamic panel data, voter group decision model.

3

1. Introduction.

As OECD member countries have recently begun to reduce their budgets,

countries will be forced to make spending decisions. Fiscal discipline will require cuts in

government expenditure leading to trade-off between different components of

government expenditure and, therefore, affecting the composition of government

expenditure. Now, in this article we explore the relationship between components of the

government expenditure and government size during the period 1970-1997 for a sample

of 26 OECD countries to shed light on how fiscal discipline might influence public

spending composition in the future.

Many authors have underscore the surprisingly little research devoted to the

determinants of the composition of government expenditure (Dunne et al. 1984, Borge

and Rattsø 1995, Sturm, 1998 and Tridimas, 2001). Moreover, Baqir (2002) claims that

most of the studies analyzing the effects of aggregate government expenditure on its

composition have focused on the economic classification of government spending.

Furthermore, those few examining the functional disaggregation of government

expenditure have concentrated in particular functions -primarily social expenditure

including education and health (Baqir, 2002) and occasionally social welfare expenditure

(Ravallion, 2002) - or in the composition of local government spending (Borge and

Rattsø, 1995). This paper innovates using the functional disaggregation of consolidated

government expenditure (COFOG, Classification of the Functions of the Government),

being as far as we are aware, the first study on the effects of fiscal consolidation in the

composition of government expenditure by functions for the OECD.

4

The assessment of the effects of aggregate government expenditure is of great

interest in the present atmosphere of budgetary cuts and cost controls. Thus, in first place,

fiscal consolidation affects economic growth through its impact on the composition of

public expenditure. Along these lines, some endogenous growth models incorporate the

composition of government spending which are capable of yielding steady state effects

(Devarajan et al. 1996, Kneller et al. 1999). Moreover, Davoodi and Zou (1998) show that

there is an optimal composition of government expenditures in which the optimal share of

each component equals its growth elasticity relative to all the function's growth

elasticities. Therefore, by changing the composition of government expenditure, fiscal

consolidation can approach or deviate the structure of public spending from its optimal

structure.

In order to investigate this aspect, section 2 reviews the existing literature on the

effects of fiscal consolidation on the composition of government expenditures. In section

3, we go on using a voter group model following Craig and Inman (1986), which lead to

stable outcomes even when issues are multidimensional. The final allocation of

government expenditure is a weighted average of each group’s preferred allocation.

Moreover, varying the allocation of government expenditure may be a process requiring a

slow adjustment. Thus, in section 4, we analyze the composition of government

expenditure in a dynamic model framework. For this purpose we will use the Generalized

Method of Moments (GMM) suggested by Arellano and Bond (1991) which, besides

specific country effects, also takes account of endogeneity of the lagged dependent

variable and of other explanatory variables. Finally, in section 5, we draw the most

significant conclusions.

5

2. Fiscal consolidation and the composition of government expenditure.

During the period 1970-1997 the share of government expenditure in GDP has

increased in the OECD from 30.5% in 1970 to 43.3% in 1997 (OECD: Economic

Outlook and National Accounts. Volume II: Detailed Tables). However, this expansion

has fluctuated in the course of the three decades. During the decade of the seventies and

the early years of the eighties the public sector increased its size constantly. This trend

was interrupted in 1983 when public expenditure as a share of GDP became stable. At the

beginning of the nineties public expenditure increased again its size until 1995 the peak

for the whole period. Thereafter, OECD countries have increasingly implement

government spending reforms towards more controlled government spending and active

deficit management (Tanzi and Schuknecht, 2000). Certainly, EU Member States signed

the Stability and Growth Pact, 1996, which constrained their public deficits and debt

levels, whereas the United States adopted the Balanced Budget Amendment. Moreover,

Zaghini (2001) shows that EU Member States have accomplished cutbacks on public

deficit through an expenditure reduction policy.

The reduction of the public sector size does not necessarily lead to proportional

decreases in the components of government spending, but may change the composition of

government expenditures by particularly affecting some of its components and protecting

others. In this way, several studies have analyzed the effects of fiscal consolidation on the

composition of government expenditures focusing on two specific components: public

investments and social spending (including education, health and social welfare

expenditure). These studies predict two different and opposite outcomes. The first strand

of studies claims that fiscal adjustments will fall over investments whereas protecting

6

social spending. Thus, Roubini and Sachs (1989) claim that at a time of fiscal

consolidation, public investments are the first to be reduced because these are the least

rigid component of expenditures. Oxley and Martin (1991) also contend that political

reasons make it easier to diminish or postpone investment spending than current

expenditure. Furthermore, Sturm (1998) suggests that myopic governments in need of

budgetary cuts reduce those less visible and long-term expenditures in order to minimize

the political costs associated with government spending cutbacks. Along these lines,

Henrekson (1988), for Sweden over the period 1950-1984, Sturm (1998), for a sample of

22 OECD countries over the period 1980-1992, and Jonakin and Stephens (1999) for a

sample of 5 Central America countries over the period 1975-1993, find that fiscal

adjustments particularly affect public investments.

Therefore, we could expect fiscal consolidation falling primarily over public

investments and protecting the rest of expenditures. Actually, studies examining the

effects of fiscal adjustments on pro-poor expenditures -mainly social expenditures-

predict that budgetary cuts will not primarily affect social spending. Thus, Ravallion

(1999) claims that if cutting expenditures save taxes to the non-poor, these voters, in turn,

would be more willing to protect pro-poor expenditure. Furthermore, Ravallion (2000)

contends that poor groups could build influential interest groups with Non-Governmental

Organizations or non-poor groups interested in avoiding the external costs of poverty.

Accordingly, Snyder and Yackovlev (2000) for a sample of 19 Latin American and

Caribbean countries during the period 1970-1996 and Cashin et al. (2001) and Baqir (2002)

for a sample of 179 and 167 countries during the period 1985-1998, respectively, find that

education and health expenditure are isolated from fiscal adjustments

7

A second strand of studies predicts that budgetary cuts will fall over social

expenditures and protect productive expenditures. Thus, Aubin et al. (1988) argue that

reducing investments –a type of productive expenditure- has more adverse political

effects than decreasing public consumption and wages because the former is more visible.

Actually, Alesina et al. (1998) show that adjustments primarily based on public transfers

and wages if anything, increased the probability of survival of governments over the

period 1960-1995 in a sample of 19 OECD countries. Moreover, Tanzi (2000) maintains

that globalization will reduce government revenues and expenditure because of the tax

competition among jurisdictions and increased mobility of factors. These authors suggest

that the reduction of government spending will not be proportional but affect particularly

to social spending and preserve productive expenditures that enhance countries

competitiveness and attractiveness to Foreign Direct Investment (FDI). Along these lines,

Keen and Marchand (1997) elaborate a model in which governments encourage country

competitiveness by raising above its optimal level the allocation to productive

expenditures and contracting utility enhancing spending such as social welfare spending.

Furthermore, Ghate and Zak (2002) elaborate a model in which politicians maximize votes

whereas voters support to politicians depending not only on the transfers they receive but

also on the of output growth. As a consequence, at a first stage social welfare expenditure

drives the growth of government, but then a threshold emerges where in order to maintain

positive output growth governments reduce aggregate government expenditure cutting social

welfare expenditure. Accordingly, Tanzi and Schuknecht (1997) show that industrialized

countries that have undertaken reforms in the size of public sector in the mid eighties, such

as New Zealand and Chile, have accomplished it through the reduction of public subsidies

and transfers. Furthermore, Ravallion (2002) finds that decreases in aggregate government

expenditure in the decade of the 80's and the 90's in Argentina has led to more than

8

proportional cuts in education, health, housing and social security spending whereas

increases in aggregate government spending did not increase at all the size of social

spending.

Among productive expenditures, Kneller et al. (1999) include those devoted to

general administration services, public order, defense, health, education, housing and

transport and communication. Therefore, these two hypothesis agree in predicting that

education and health expenditures –considered both, as social and productive spending-

would be protected from fiscal adjustments but disagree in the effects of fiscal

consolidation on public investments –at least those related to transport and

communications- and social welfare spending.

3. Theoretical model.

The base model under which the determinants of government expenditure are

usually considered is the median voter model, developed from the studies of Borcherding

and Deacon (1972) and Bergstrom and Goodman (1973). In this model it is assumed that

citizens vote by means of a majority system, voter preferences are single-peaked and the

size of the public sector is the single issue to be decided. However, Craig and Inman

(1986) show that when issues are multidimensional, as the composition of government

expenditure, majority rule does not lead to a stable equilibrium because any allocation

can be beaten by another proposal, cycling from one proposal to another. It is required to

add legislative structures to the majority rule to insure a stable allocation. Along these

lines Craig and Inman (1986) propose a voter group decision model, where the final

outcome, the structure-induced equilibrium, is a weighted average of each group’s

preferred allocation. Following Borge and Rattsø (1995), we use Craig and Inman’s

9

model assuming that aggregate government expenditure is already determined when voter

groups choose the composition of government expenditure. This assumption is also usual

in the literature analyzing the composition of government expenditure or some of its

components (Dunne et al. 1984, Sturm, 1998 and Baqir, 2002) including those adopting a

consumer demand framework (see Tridimas, 2001 for a survey). Therefore, in this model

voter groups and not governments decide which components are primarily to be reduced

(increased) at times of fiscal contractions (expansions). Thus, we assume that a group h

maximizes a utility function that depends on the per capita consumption of component f

of government expenditures (G1,..,Gf,..,GF), per capita private consumption (Yh) and

structural characteristics (Z), subject to a budget constraint such:

The inclusion of structural characteristics in the utility function still makes that the

problem has the form of a consumer problem where G is the previously determined

aggregate government expenditure and the price for component of government

expenditure Gf is Pf = Cf N(ηf -1), where Cf is the cost of a unit of public good and service

of component f, N is total population and ηf the degree of congestion of component f of

public goods and services1. Per capita private consumption Yh is taken into account to

capture the degree of complementarily or substitutability between government services and

private consumption. The final outcome will be a weighted average of each group’s

preferred allocation:

1 As Borcherding and Deacon (1972) we assume that there is no tax discrimination, so that the tax share T

for each individual is T = 1/N. This assumption does not introduce a significant bias in the empirical

estimation, as it seems reasonable to assume that changes in price tax are inversed to changes in population

(Borcherding et al. 2003 ).

( ) )1(.,....,1;..,,,....,max1

1 FfGGPtsZYGGUF

fff

hF

h ==∑=

10

Where ωh is the political strength of group h. We assume that this political

strength is proportional to the share of each voter group in total population2. Groups will

be differentiated by the age structure of the population, considering three groups:

population below 15 years, over 64 and between 15 and 64. Indeed, age appears to be as

one of the most important characteristics shaping voter’s preferred allocations. Thus, age

groups may have a higher demand for those components of government expenditures for

which they are the primarily beneficiaries. In fact, Gemmell et al. (1999) suggest that the

omission of the age structure of the population may give rise to bias in the estimation of

the rest of elasticities. Due to lack of data availability we assume that unit costs across

components are equal (Cf=C∀ f) and that each age group has the same income per capita.

Therefore we cannot estimate cross-price elasticities and constrain income elasticity to be

the same across age groups. On the other hand, we enable the public/private sector price

ratio to be modified in the course of time, including relative prices (Pr=C/Px, where Px is

the price of the private sector, Gemmell et al., 1999 and Tridimas, 2001). Replacing Pf

and Gf in (2), and assuming constant elasticities, we could finally express the demand

system as:

Where Gf* is government spending in component f in real terms with Gf=Gf

*N-ηf,

G* is aggregate government spending with G=G*N-η and αf, βf, φf, φh,f and ωf are income,

price, total population, age group and aggregate government expenditure elasticities of

2 Nevertheless increases in the age group may not necessarily lead to increases in the political strength

because it may also raise the incentive to free ride of the former members of the group (Olson, 1965).

( ) )2(1.,....,1;,...,1;,,,,,...,11

1 ∑∑==

====

H

h

hH

h

hF

hhf HhFfZNGYPPfG

Gωω

( ) ( )( )

(3);,...,1

1211;*3

1*

* ,

Ff

GNN

NPYG

Gfffff

h

hr

f ffh

fff

=

−+−+−+=

=

∏

=

ηωηηηβφω

φφβα

11

component f of government expenditures and h group in the age interval 0-14 years, 15-

64 years and over 64, respectively. Protected components will have a government

elasticity (wf) negative indicating that the share of this type of spending will increase

(decrease) as aggregate government decrease (increase). Fiscal consolidations will fall

over those expenditures with positive government elasticity, whereas those expenditures

with government elasticity not significantly different from zero would be reduced or

increased proportionally to changes in aggregate government expenditures. The model

does not explain why fiscal consolidations happen at all, as aggregate government

expenditure is assumed to be previously determined, but focus on the effects of aggregate

government expenditure changes on its composition. The model is consistent with the

strand of studies that argue that governments reduce those expenditures associated with

less political costs. Thus, voter groups decide which component’s cutback they prefer

more or they dislike less. Hence, governments seeking to minimize the political costs of

reducing aggregate government expenditure would just cutback those components

decided by voter groups. Our model would be only consistent with Keen and Marchand

(1997), Tanzi (2000) and Tanzi and Schuknecht (2000) if governments reduce social

spending and protect productive expenditures because this happens to be the decision of

voter groups and not juts the strategy of governments for encouraging the

competitiveness and attractiveness to FDI of the country.

We also introduce dynamics in the model to capture that adapting to changes of

voter group’s demand or changes of group’s political strength is a process requiring a slow

adjustment. Actually, some of the government expenditures show a high degree of rigidity

since some of the expenditure are previously committed, as those related to social security

and public employees wages. Introducing dynamics we also match Craig and Inman’s

12

(1986) model in that the final outcome depends not only of the group’s preferred

allocation but also on the status quo, reflected by the composition of government

expenditure in the previous year. Actually, the incrementalist decision-making literature

has shown the decisive advantage of the status quo in the determination of government

expenditure (Borge and Rattsø, 1995). Finally, the model will be used for explaining

consolidated government expenditures for a panel of 26 OECD countries over the period

1970-1997 (all State Members of the OECD except Poland, Czech Republic, Hungary and

Slovak Republic). We consider nine components of government expenditures corresponding

to the functional disaggregation of government expenditure. Components are arranged in a

very similar order to the one used in the Classification of Functions of the Government

(COFOG, United Nations, 2000): public services (including general administrative

services and public order and law), defense, health, education, social welfare, housing,

economic services (making a distinction between transport and communications and other

economic services), and recreational, cultural and religious affairs. We exclude other non-

classified functions (mainly interest payments) that are exogenously determined.

Therefore, the share of a component is measured as the proportion of this component in

the aggregate government expenditure, excluding interest payments. To sum up we have

a system with demand functions for F-1=8 components of government expenditure and

inferring coefficients for recreational, cultural and religious affairs from the budget

restriction. Thus, assuming that actual composition of government expenditure does not

match the voter groups’ desires immediately, but by a partial adjustment process, expressing

the demand functions in logarithmic form and rearranging terms, we get

13

( ) ( ) ( ) ( )

( ) ( )

)4(26,....1:;1997,...1970:.8,....1:

lnlnlnln

lnlnlnln1ln

,

1

1,

*

,

,,3,3

,

,,1,1,

,,,,

*1,

*

*,

*,,

,

1,,

ktf

uY

G

N

N

N

NN

PYa

a

G

G

G

G

tk

K

kk

tkf

tk

tkf

tk

tkftkff

tkrftkffk

kf

tktk

tkf

tk

tkf

λελλωλφφλωφλ

λβωαλλλ

++

+

+

+++

++++

+

−=

∑−

=

−

−

Where λ is the constant speed of adjustment, k is the country, t is the year, uk is a

dummy than takes value 1 for country k and 0 otherwise, and εt,k is the classical disturbance

term3.

4. Data and econometric analysis.

Data for government expenditures is built on OECD publication National Accounts.

Volume II: Detailed Tables. This source is chosen inasmuch as it offers information on the

consolidated spending of all levels of government and, in addition, it follows the accrual

criterion4. Relative prices are approximated as in Gemmell et al. (1999) by the ratio of the

public sector deflator to the GDP deflator. Public sector deflator is the result of the weighted

mean of government investments deflator, public consumption deflator and public transfers,

the latter represented by the consumer price index, all obtained from OECD: Economic

Outlook. The per capita income (in Purchasing Power Parities of the 1995 dollar and in real

terms of that year), population and government expenditures series are obtained from the

3 We focus on age groups 1 and 3, the share of young and old population, since these groups do have

specifically needs for public goods and services.

4Data from national agencies, OECD and World Bank country reports, Eurostat: General Government Accounts

and Statistics and the IMF publication: Government Finance Statistics, is used on a supplementary basis so as

to obtain longer statistical series. Although IMF data covers a longer period of time, it is not consolidated for

all levels of public administrations as a rule and uses the cash criterion.

14

OECD: National Accounts: Volume I. Main Aggregates, whereas the age structure of the

population is taken from the OECD: Labor Force Statistics. We compute permanent

income per capita since demand is based on permanent income rather that on temporary

income levels (Peltzman, 1980). We approximate permanent income per capita by taking

a three-year moving average reducing the sample by two observations for each country.

We use the GMM estimator suggested by Arellano and Bond (1991), based on

taking first differences on (4), dropping unobservable country dummies. However, taking

first differences introduced bias because of the correlation between ∆εk,t and ∆ln(G*f,k,t-1/

G*k,t-1). Thus we use as instruments for ∆ln(G*f,k,t-1/G*k,t-1) at least two periods lagged values

of ln(G*f,k,t/ G*k,t). Moreover, income per capita introduces simultaneity as government

expenditure and its composition affect long run economic growth5. Furthermore, there

might also be correlation between the equation errors and aggregate government spending.

Therefore, we take as instruments for the first differences of income per capita and the size

of aggregate government expenditure at least two lagged values of the levels of income per

capita and the size of aggregate government expenditures. Finally, note that after taking first

differences we introduce negative first order autocorrelation in the transformed model,

which would be second order serial correlation if in the original model there was already

first order serial correlation. In this latter case we would have to use instruments at least

three periods lagged values of the levels of income per capita and the size of aggregate

government expenditure. As it can be seen from Table 1, the M2 tests do not reject the null

5 See Agell et al., 1999, for a survey on fiscal policy and economic growth. This endogenous issue has not

received much attention in empirical studies (Borcherding et al. 2003). This may be due to the fact that this

potential source of endogeinity is reduced considering that the size and composition of public spending

effects on economic growth are likely to unfold slowly (Devarajan et al., 1996).

15

hypothesis of absence of second-order serial correlation. Furthermore, the Sargan test

statistic of overidentifying restrictions does not reject the validity of the instruments used.

[Table 1, around here]

Focusing on the variable of interest, we find a negative and significant relationship

between aggregate government expenditure and education, health and social welfare

expenditure. Therefore, reductions (increases) in aggregate government expenditure increase

(reduce) the share devoted to social expenditures. Thus we find some indication of social

expenditure being the most isolated both from fiscal contractions or expansions. This result

is in line with Dunne et al. (1984) for general current government expenditure in the UK

during the period 1950-1980 and Borge and Rattsø (1995) for local expenditure in Norway

during the period 1986-1989, who find that education expenditure have the property of

necessities. Furthermore, our results are consistent with Snyder and Yackovlev (2000), who

find that primary and secondary education spending are relatively more insulated from

economic shocks and Cashin et al. (2001) and Baqir (2002), who find that at a time of

budgetary cuts, not only education but also health expenditure is protected. On the other

hand, public services, housing and economic services show a positive and significant

relationship with aggregate government expenditures, giving indication of more than

proportional reactions to changes of the size of aggregate government expenditure. Hence, it

seems that at a time of fiscal contractions, budgetary cuts fall over housing, economic

services and public services expenditures. Finally, we find that transport and

communications and defense react proportionally to changes in the size of government

expenditure.

These results do not entirely support or reject any of the two hypotheses considered.

16

Thus, we find that education and health expenditure are protected from fiscal cycles, as

claimed by both hypotheses. Nevertheless, we do not find any indication of transport and

communications, the most productive expenditure (Easterly and Rebelo, 1993) and the

most associated with public investment, being particularly protected or targeted at times

of fiscal adjustment, which is in contrast with both hypotheses. Finally, we find that social

welfare expenditure is the most isolated spending from fiscal contractions or expansions as

predicted by the first hypothesis and in sharp contrast to the second.

The rest of results are in line with the economic literature. From the speed of

adjustment, it can be seen that education and defense expenditure seem to be the most rigid

of the components of government expenditure. These results are in line with Dunne et al.

(1984), who find that military expenditures is the component showing the most long planed

expenditure character among general current government expenditure in the UK over the

period 1950-1980 and Golini (1998) who suggest that education have a low speed of

adjustment because most of the staff is pre-committed for long term and schools are also a

fixed stock. The effect of income changes increases the share of functions such as education,

social welfare, public services and transport and communications, confirming that Wagner's

law is especially applicable for welfare services (Peacock and Scott, 2000). Social welfare,

health, and education are the least price elastic, a result consistent with Baumol's conjecture

(1967), since wage and salary predominant in health and education expenditure. Population

reduces the allocation to pure public goods such as defense (Murdoch and Sandler, 1985

for Australia military demand over the period 1961-1979), transport and communications

(Randolph et al., 1996, for 27 low and middle-income countries over the period 1980-1996)

and economic services whereas increases merit good services as education and housing

along with social welfare. Elderly population increases the demand for health and social

17

security as this age group receives more benefits from it than the others age groups. This

result is partially consistent with Lindert (1996), who finds that ageing population is

strongly and positively associated with social welfare spending for a panel of 19 OECD

countries during the period 1960-1981. Nevertheless, this author does not find any

significant effect of elderly population on health expenditure. We also find evidence of

elderly reducing the share of education in total government expenditure corroborating the

findings of Bergstrom et al. (1982) for individual survey data in Michigan in 1978, Borge

and Ratsso (1995) for local public spending in Norway during the period 1986-1989, and

Fernandez and Rogerson (2001) for state spending in the U.S. over the period 1950-1990.

As for young population we find that this group significantly increases the size of social

security, showing evidence of the relevance of family and child benefits in the social

welfare spending (Than Dang et al, 2001), whereas it does only significantly decrease

defense expenditure. Finally, we find a positive elasticity of young population on the share

of education but not to a significantly level.

5. Asymmetry in fiscal contractions and expansions.

In previous section we have constrained the effects of fiscal expansions and

contractions on the composition of government expenditure to be the same. Nevertheless,

components may react differently to increases or decreases in government spending.

Actually, Ravallion (2002) finds that social spending in Argentina during the 80’s and 90’s

decade decreased more than proportionally after reductions in government expenditure

whereas it did not significantly increase following raises in government spending. Thus we

include a variable in the transformed model, which takes the value of the first differences if

government expenditure is reduced and zero otherwise. Table 2 shows results of the

18

estimation taking into account possible asymmetry on the effects of government size

changes. Again, M2 do not reject the null hypothesis of absence of second-order serial

correlation. Furthermore, the Sargan test statistic of overidentifying restrictions does not

reject the validity of the instruments used.

[Table 2, around here]

Looking at the coefficients associated to government size it can be seen that some of

the components of government expenditure have a significantly different pattern if

government size is increased or if it is reduced. Thus, we find that the share of defense in

aggregate government spending is reduced either when government size is increased or

decreased. Therefore we find indication that fiscal consolidations fall over defense along

with economic services. This result is in line with Gupta et al. (2001), who find that fiscal

consolidations fall primarily over defense expenditure in a sample of 120 countries over the

period 1985-1998, and in contrast to Davoodi et al. (1999) and Jonakin and Stephens (1999),

who find that defense expenditure is more protected when fiscal discipline is implemented in

a sample of 130 countries for the period 1985-1998 and 5 Central America countries over

the period 1975-1993 respectively. Public services and housing increase their shares in

aggregate government expenditures in fiscal expansions, and to some extent, also in fiscal

contractions. Finally, some of the components show the same sign associated with increases

or decreases of government size, but with different magnitudes. Interestingly, the share of

social welfare expenditure decreases in fiscal expansions by a higher magnitude than the

increases associated with fiscal contractions. In contrast, the share of education is more

isolated from fiscal contractions than from fiscal expansions. Actually, we find evidence of

social welfare expenditure being preserved during fiscal contractions but to a lesser extent

than education and health expenditures. For the rest of variables we find coefficients

19

close to the table 1, though standard errors generally increase.

To sum up, we find evidence of social spending being isolated from fiscal

expansions or contractions but to different degrees. Thus, social welfare reduces its share in

the aggregate government expenditure more than education and health spending in fiscal

expansions, whereas these two latter expenditures increase their shares more than the former

in fiscal contractions. In fact, education is the most protected component of government

expenditure. Therefore, social spending with a productive character –education and health

spending- is more protected than social spending without productive character –social

welfare spending.

6. Conclusions.

This paper explores how fiscal consolidation can affect the composition of

government expenditures analyzing the relationship between the size of aggregate

government expenditure and each of its functional components in the OECD over the

period 1970-1997. For this purpose we employ a voter group model following Craig and

Inman (1986) in which the final allocation is a weighted average of the preferred

allocations of the groups considered. We differentiate between young, middle-aged and

elderly population, as age appears to be the most important characteristic shaping voter’s

preferred allocations. Furthermore, we introduce dynamics in the model because changes in

the composition of government expenditure require a slow adjustment process. Previous

economic literature predicted two contradictory effects of fiscal consolidation on the

composition of government expenditure. The first strand of literature predicts that

governments will reduce public investment because decreasing social spending has more

20

political costs. On the other hand, the second strand of literature predicts that

governments will protect those more productive expenditures and make the fiscal

consolidation fall over social welfare spending to enhance the competitiveness and

attractiveness to FDI of the country. In our set up age groups and not governments choose

the composition of government expenditure following changes in the size of government,

which is assumed to be previously determined.

Thus, we find that at times of fiscal consolidation, expenditures with a social

character are the most protected. But the productive character of the expenditure also

matters, because among social expenditures, those being also productive such as

education and health are even more protected than those without productive character as

social welfare. Moreover, fiscal consolidation does not fall primarily over the most

productive expenditure, transport and communications, but over defense and economic

services. Hence, it seems as if voter groups protect first social spending, and among these

expenditure, care specially about those with a productive character. This result may be

given some evidence of voter groups realizing that reducing productive expenditures

harm long-term economic growth. Therefore, voter groups reach an equilibrium between

utility and economic-growth enhancing expenditure by protecting social spending with a

productive character the most. Voters may also take into account the growth effects when

deciding the composition of government expenditure. Along these lines, Ghate and Zak’s

(2002) elaborate a model in which voters support to politicians depend not only on the

transfers they receive but also on the of output growth. Voters may be willing to accept

some reductions on the transfers they desired if they think that this stimulates output

growth. Nevertheless, these results are also consistent with governments reducing pure

public goods, such as defense and economic services, whereas protecting merit goods,

21

such as education and health expenditure, and transfers when facing budgetary cuts. As

pure public goods are less visible than merit goods, by reducing the former, governments

may be minimizing the political costs of fiscal consolidations.

Finally, we show that the effects of fiscal consolidation on government expenditure

components depend firstly on the purpose of the expenditure, and secondly on its productive

character. Therefore, it is more useful to analyze the composition of government

expenditure by purpose, using its functional classification rather than the economic

classification. This is an important innovation of this paper, since most of the few analysis

on the effects of fiscal consolidation on the composition of government expenditures have

focused on its economic classification.

22

References

Alesina, Alberto, Perotti, Roberto, Tavares, Jose, 1998. The Political Economy of

Fiscal Adjustments. Brokings Papers on Economic Activity Spring, 197-266.

Arellano, Manuel, Bond, Stephen, 1991. Some Tests of Specification for Panel

Data: Monte Carlo Evidence and an Application to Employment Equations. Review of

Economic Studies 58, 277-297.

Aubin, Christian, Berdot, Jean P., Goyeau, Daniel, Lafay Jean D., 1988. The

growth of public expenditure in France. In: Lybeck, Johan A., Henrekson, Magnus (Eds.),

Explaining the growth of government. Elsevier Science Publishers, Amsterdam, pp.201-

230.

Baumol, William J., 1967. Macroeconomics of Unbalanced Growth: The

Anatomy of Urban Crises. American Economic Review 57, 415-26.

Baqir, Reza, 2002. Social Sector Spending in a Panel of Countries. Working Paper

WP/02/35. International Monetary Fund, Washington.

Bergstrom, Theodore C., Goodman, Robert P., 1973. Private demands for public

goods. American Economic Review 63, 280-296.

Bergstrom, Theodore C., Rubinfeld, Daniel L., Shapiro, Perry, 1982. Micro-Based

Estimates of Demand Functions for Local School Expenditures. Econometrica 50, 1183-

1205.

Borcherding, Thomas E., Deacon, Robert T., 1972. The demand for the services

of non-federal governments. American Economic Review 62, 891-901.

Borcherding, Thomas E., Ferris, Stephen F., Garzoni, Andrea, 2003. Growth in

the Real Size of Government since 1970. In: Wagner, Richard E., Backhaus, Jürgen G.,

(Eds.), Handbook of Public Finance. Kluwer Academic Publisher, Dordrecht,

23

forthcoming.

Borge, Lars E., Rattsø, Jørn, 1995. Demographic Shift, Relative Costs and the

Allocation of Local Government Consumption in Norway. Regional Science and Urban

Economics 25, 705-726.

Cashin, Paul A., Mauro, Paolo, Pattillo, Catherine A., Sahay, Ratna, 2001.

Macroeconomic Policies and Poverty Reduction: Stylized Facts and an Overview of

Research. Working Paper 01/135, International Monetary Fund, Washington.

Craig, Steven G., Inman, Robert P., 1986. Education, Welfare and the ‘new’

Federalism: State Budgeting in a Federalist Public Economy. In: Rosen, Harvey S. (Ed.),

Studies in State and Local Public Finance. Chicago University Press, Chicago, pp. 187-

222.

Davoodi, Hamid, Zou, Heng-fu, 1998. Fiscal decentralization and economic

growth: A cross country study. Journal of Urban Economics 43, 244-257.

Davoodi, Hamid, Clements, Benedict, Schiff, Jerald, Debaere, Peter, 1999.

Military Spending, the Peace Dividend, and Fiscal Adjustment. Working Paper,

WP/99/87, International Monetary Fund.

Devarajan, Shantayanan, Swaroop, Vinaya, Zou, Heng-fu, 1996. The

Composition of Public Expenditure and Economic Growth. Journal of Monetary

Economics 37, 313-344.

Dunne, John P., Pashardes, Panos, Smith, Ronald P., 1984. Needs, Costs and

Bureaucracy: The Allocation of Public Consumption in the UK. Economic Journal 94, 1-

15.

Easterly, William, Rebelo, Sergio, 1993. Fiscal Policy and Economic Growth.

Journal of Monetary Economic 32, 417-458.

Fernandez, Raquel, Rogerson, Richard, 2001. The Determinants of Public

24

Education Expenditures: Longer-Run Evidence from the States, 1950-1990. Journal of

Education Finance 27, 567-583.

Gemmell, Norman, Morrissey, Oliver, Pinar, Abuzer, 1999. Fiscal illusion and the

demand for government expenditures in the UK. European Journal of Political Economy

15, 687-704.

Ghate, Chetan, Zak, Paul J., 2002, Growth of Government and the Politics of

Fiscal Policy, Structural Change and Economic Dynamics 13, 435-455.

Golini, Antonio, 1998. Demographics, Public Spending and Social Programs.

Labour 12, 119-143.

Gupta, Sanjeev, Mello, Luiz, Sharan, Raju, 2001. Corruption and military

spending. European Journal of Political Economy 17, 749-777.

Henrekson, Magnus, 1988. Swedish government growth. A disequilibrium

analysis. In: Lybeck, Johan A., Henrekson, Magnus (Eds.), Explaining the growth of

government. Elsevier Science Publishers, Amsterdam, pp. 93-132.

Jonakin, Jon, Stephens, Mark, 1999. The impact of adjustment and stabilization

policies on infrastructure spending in Central America. North American Journal of

Economics and Finance 10, 293-308.

Keen, Michael, Marchand, Maurice, 1997. Fiscal competition and the pattern of

public spending. Journal of Public Economics 66, 33-53.

Kneller, Richard, Bleaney, Michael and Gemmell, Norman, 1999. Fiscal Policy

and Growth: Evidence from OECD Countries. Journal of Public Economics 74, 171-190.

Lindert, Peter H., 1996. What Limits Social Spending?. Explorations in Economic

History 33, 1-34.

Murdoch, James C., Sandler, Todd, 1985. Australian Demand for Military

Expenditures: 1961-1979. Australian Economic Papers 24, 142-153.

25

Olson, Mancur, 1965. The Logic of Collective Action. Harvard University Press,

Cambridge.

Oxley, Howard, Martin John P., 1991. Controlling government spending and

deficit: trends in the 1980s and prospects for the 1990s. OECD Economic Studies 17,

145-189.

Peacock, Alan T., Scott, Alex, 2000. The curious attraction of Wagner's law.

Public Choice 102, 1-17.

Peltzman, Sam, 1980. The Growth of Government. Journal of Law and

Economics 23, 209-288.

Randolph, Susan, Bogetic, Zeljko, Hefley, Dennis, 1996. Determinants of Public

Expenditure on Infrastructure. Transportation and Communication. Policy Research

Working Paper 1661. World Bank, Washington.

Ravallion, Martin, 1999. Is More Targeting Consistent with Less Spending?.

International Tax and Public Finance 6, 411-419.

Ravallion, Martin, 2000. Are the Poor Protected from Budget Cuts? Theory and

Evidence for Argentina. Policy Research Working Paper 2391, World Bank, Washington.

Ravallion, Martin, 2002. Are the Poor Protected from Budget Cuts? Evidence for

Argentina. Journal of Applied Economics V, 95-121.

Roubini, Nouriel, Sachs, Jeffrey, 1989. Government spending and budget deficits

in the industrial countries, Economic Policy 8, 99-132.

Snyder, James M., Yackovlev, Irene, 2000. Political and Economic Determinants

of Government Spending on Social Protection Programs. Working Paper. Massachusetts

Institute of Technology, Cambridge.

Sturm, Jan E., 1998. Public Capital Expenditure in the OECD Countries: the

Causes and Impact of the Decline in Public Capital Spending, Edward Elgar Publishing,

26

Cheltenham.

Tanzi, Vito, Schuknecht, Ludger, 1997. Reforming government: An overview of

recent experience. European Journal of Political Economy 13, 395-417.

Tanzi, Vito, 2000. Globalisation and the Future of Social Protection. Working

Paper No. 12. International Monetary Fund, Washington.

Tanzi, Vito and Schuknecht, Ludger, 2000. Public Spending in the 20th Century:

A Global Perspective. Cambridge University Press, Cambridge.

Than Dang, Thai, Antolin, Pablo, Oxley, Howard, 2001. Fiscal implications of

ageing: projections of age-related spending. OECD Economics Department Working

Paper No. 305. OECD, Paris.

Tridimas, George, 2001. The economics and politics of the structure of public

expenditure. Public Choice 106, 299-316.

Zaghini, Andrea, 2001. Fiscal Adjustments and economic performing: a

comparative study. Applied Economics 33, 613-624.

27

Table 1: Estimation results of the determinants of the composition of government expenditure (GMM, Two-step estimate, Arellano and Bond, 1991)

Share in aggregate government expenditure of a: Dependent variable Public

Services Defense Health Education Housing Trans. &

Comm. Economic services

Social Welfare

Rec. & cult affairs

Lagged –1 Gf,t-1

0.398 (0.00)

0.615 (0.00)

0.427 (0.00)

0.835 (0.00)

0.249 (0.00)

0.439 (0.00)

0.492 (0.00)

0.334 (0.00)

-

Government Size Gt

0.514 (0.00)

-0.230 (0.13)

-0.422 (0.00)

-0.153 (0.04)

1.139 (0.00)

0.123 (0.26)

0.707 (0.00)

-0.355 (0.00)

1.033 (0.55)

Income Y

0.927 (0.03)

-0.703 (0.02)

0.009 (0.95)

0.387 (0.05)

-1.286 (0.03)

0.580 (0.10)

0.121 (0.78)

0.270 (0.00)

-10.028 (0.02)

Relative Prices Pr

-1.259 (0.00)

-0.415 (0.31)

0.821 (0.00)

0.327 (0.09)

-0.819 (0.01)

-0.477 (0.28)

-0.507 (0.01)

0.870 (0.00)

-7.734 (0.04)

Population N

-0.086 (0.86)

-0.743 (0.01)

0.060 (0.78)

0.417 (0.05)

2.494 (0.00)

-1.495 (0.00)

-2.311 (0.00)

1.223 (0.00)

-6.326 (0.24)

% Pop < 15 N1/N 0.246 (0.53)

-1.307 (0.00)

0.062 (0.76)

0.309 (0.16)

-0.667 (0.15)

0.420 (0.12)

0.056 (0.95)

0.470 (0.00)

-7.212 (0.09)

% Pop >64 N3/N -1.034 (0.05)

-0.121 (0.70)

0.436 (0.02)

-0.611 (0.02)

-0.854 (0.44)

-0.422 (0.45)

0.016 (0.98)

0.371 (0.00)

6.390 (0.23)

M1 (26)

-1.35 (0.17)

-2.09 (0.04)

-2.40 (0.02)

-2.87 (0.00)

-1.43 (0.15)

-3.080 (0.00)

-2.49 (0.01)

-1.001 (0.31)

-

M2 (26)

0.34 (0.73)

0.90 (0.36)

0.10 (0.92)

-0.53 (0.60)

0.46 (0.64)

-1.345 (0.18)

-1.49 (0.14)

-0.981 (0.33)

-

Sargan test

18.85

21.00

24.48

20.85

14.88

20.21

20.97

22.82

-

Degrees of freedom

48

43

46

67

68

43

66

44

-

Sample: 26 OECD Countries (All actually OECD Member countries, except Poland, Czech Republic, Hungary and Slovak Republic), 1971-1996.

a p-values in parentheses

28

Table 2: Estimation results distinguishing between increases and decreases of aggregate government expenditure (GMM, Two-step estimate, Arellano and Bond, 1991)

Share in aggregate government expenditure of a:

Dependent variable Public Services

Defense Health Education Housing Trans. & Comm.

Economic services

Social Welfare

Rec. & cult affairs

Lagged –1 Gf,t-1

0.293 (0.00)

0.624 (0.00)

0.283 (0.00)

0.845 (0.00)

0.150 (0.00)

0.345 (0.00)

0.381 (0.00)

0.323 (0.00)

-

Government Size Gt

1.092 (0.00)

-0.311 (0.07)

-0.254 (0.05)

-0.102 (0.07)

1.917 (0.01)

-0.042 (0.83)

0.409 (0.09)

-0.615 (0.00)

1.458 (0.54)

Government Size Decreases

-1.185 (0.05)

0.383 (0.02)

-0.457 (0.18)

-0.369 (0.01)

-2.141 (0.04)

-0.200 (0.56)

0.457 (0.36)

0.463 (0.00)

5.843 (0.28)

Income Y

1.174 (0.03)

-0.264 (0.50)

-0.066 (0.79)

0.084 (0.69)

-3.026 (0.00)

0.446 (0.54)

0.246 (0.59)

0.300 (0.04)

-7.701 (0.12)

Relative Prices Pr

-1.561 (0.00)

-0.421 (0.40)

0.947 (0.00)

0.606 (0.00)

-0.725 (0.03)

-0.213 (0.70)

-0.471 (0.03)

0.850 (0.00)

-8.638 (0.04)

Population N

-1.456 (0.10)

-0.400 (0.46)

-0.512 (0.25)

-0.064 (0.79)

1.794 (0.46)

-2.210 (0.01)

-1.748 (0.14)

1.924 (0.00)

-2.681 (0.80)

% Pop < 15 N1/N 0.667 (0.22)

-1.016 (0.07)

0.248 (0.27)

0.123 (0.53)

-1.115 (0.08)

0.270 (0.68)

-0.389 (0.74)

0.321 (0.05)

-4.449 (0.42)

% Pop >64 N3/N -1.994 (0.04)

-0.339 (0.26)

0.461 (0.11)

-0.494 (0.03)

-1.087 (0.29)

-0.264 (0.71)

-0.441 (0.71)

0.461 (0.00)

10.918 (0.03)

M1 (26)

-1.13 (0.26)

-1.01 (0.31)

-1.01 (0.31)

-2.85 (0.00)

-0.88 (0.38)

-2.45 (0.01)

-2.26 (0.02)

-1.01 (0.31)

-

M2 (26)

-0.04 (0.98)

-0.99 (0.32)

-0.99 (0.32)

-0.38 (0.70)

0.49 (0.63)

-1.45 (0.15)

-0.54 (0.12)

-0.99 (0.32)

-

Sargan test

17.21

22.86

24.09

19.87

14.52

19.56

19.90

23.07

-

Degrees of freedom

46

42

45

66

67

46

65

44

-

Sample: 26 OECD Countries (All actually OECD Member countries, except Poland, Czech Republic, Hungary and Slovak Republic), 1971-1996.

a p-values in parentheses